Michael Dell——Invest America法案正式成法、AI人才争夺、算力需求与市场更新｜BG2

Are the productivity gains from this going to be as big or bigger than what we saw from personal computers and the internet?

Oh, it's far bigger. It's far, far bigger. Yeah, I feel 98% confident.

Hey guys, great to see you both. Bill, maybe you're up in Tahoe or something, and we're thrilled to have one of our great friends, Michael Dell, on to chop it up with us. Happy 4th of July, you guys.

Most of you know Michael. He's built one of the most iconic technology companies, starting in his dorm room 41 years ago, Michael. I think you just had your 40th anniversary.

41 years ago.

41 years ago. Now you're a major player in AI. You spun off VMware, of course, and now you're a major owner of Broadcom. Dell remains a $100 billion business. I think you own a lot of the business—maybe half of it—and it's one of the largest builders of AI servers on the planet.

In addition to that, you and your incredible wife, Susan, have an amazing foundation. You do great work in Texas and around the country. I saw that you just contributed to disaster relief. What a tragedy in Texas. Kudos to you both for all the good work you do on so many issues, but particularly in the state of Texas.

It was great to have you and Susan as partners on Invest America. I know we're going to talk about that today. You wrote the book Play Nice But Win. I understand that reading an entire book is a very painful process for an author, but with your voice and inflection, you get nuances I don't think you would get just from the writing.

Well, thank you for saying that, Bill, and I appreciate the kind introduction, guys. Look, I mean, I think if you're going to take the time to write a book, which is a major endeavor if you really do it yourself and do it right—and I did have somebody helping me, by the way, so I'm not going to take full credit for it—I think you should take the additional time to record the audiobook because you can display emotion and intonation and really tell the story in your own voice. It's a powerful way to convey thoughts and emotions at the same time.

I love audiobooks. I love going outside and walking and hiking, and turning on a good audiobook is a great way to do it.

Totally. Well, in the spirit of storytelling, Michael, do you remember when you first met Gurley?

Yeah, I do remember when I first met Gurley. This was in the '90s, and Bill had written this research report that was super thick. I'm reading this report, and I'm like, how the bleep does this guy know more about our business than we do? It's like, what? We must be totally screwing up here.

He had uncovered a whole bunch of analysis and thoughts about our business, and we were so busy and distracted by growth that we had missed a few things. Bill shined a massive light on that, and it was super helpful. So I became a fan instantly of his work, and I've been a fan ever since.

Brad, that was 32 years ago. Even when I read the report, I was like, “Damn, we should have figured this out.”

This was 32 years ago. I was 28, Michael was 29, and we were running a public company. I've always cherished the fact that Michael's a year older than me, so I always have time to catch up.

1. Lessons from the 90’s at Dell

Well, it's great to be here with the kids. Bill, tell us just a second about the '90s, because I think it has some parallels to the period that we're living in now.

Dell was growing incredibly fast, obviously, building low-cost, high-quality computers. What inspired you to start covering Dell, and what led you to these insights? Did you just focus on that company? Was this a breakthrough piece of work for you?

Well, Michael's heard this before, so I hopefully won't bore him, but I had worked in the PC industry. I spent over 2 years at Compaq in Houston. Interestingly, I think being inside of Compaq, we had a view of Dell that wasn't as respectful as it should have been.

Once I got outside and was able to look at the numbers in a different way, I was able to see things more clearly. But the gentleman who made this all click for me was Michael Mauboussin, who you know. He had taught me to look at return on invested capital. That's part of what Michael was referring to.

The company had insane balance-sheet turnover, in a way that the cash flow relative to the earnings was really high, and the ROIC was 10x anyone else in the business. For some reason, probably just youth, I went and did a strong buy on the initiation, which, Michael, made a bunch of his employees rich. I ended up making a bunch of the salespeople rich there at CSFB, just as a result of riding on their coattails. It was quite a time.

The '90s were fun. The '90s were fun. The stock went up 130,000%. We had 7 stock splits.

What was the value of the company when you went public?

What was the total enterprise value or market cap when you went public, Michael?

It was like $150 million or something like that.

I mean, see, that's the beautiful thing. That's like a Series A in venture capital.

It went up 100x after this initiation in the public markets.

Yeah, Bill, I just need you to recommend our stock one more time, and then—

There was an element that I think is super interesting, and that was also part of what Michael was referring to: their inventory turns were so damn high compared to the rest of the industry. They were building to individual customer orders. They weren't building to inventory; they were building to demand.

Because component prices fell so much, we calculated they got a 200-basis-point gross-margin advantage just by having the FIFO queue.

Oh, my goodness. Yeah, exactly. Just in time. This was a structural competitive advantage. By the way, it still is.

The point is that the cost of the materials is always coming down. If your competitor has, let's say, 90 days of inventory in a series of queues with distributors and dealers, and you have 6 days of inventory—which we actually had for about 7 years in a row—6 days of inventory, think about that.

It's a structural competitive advantage because you have fresher inventory and fresher costs.

And, of course, you don't have all that capital tied up. Your return on capital is essentially infinite, especially when you're paying your suppliers for a period longer than your customers pay you. So you have a negative cash-conversion cycle.

Which we still have, typically around a negative 50-day cash-conversion cycle.

That's powerful.

It's a beautiful thing.

Yes, it is. Well, let's transition from 32 years ago to the present.

Cash is king. Everything else is an opinion.

2. Invest America Act Signed Into Law

No doubt. No doubt. So, you two just had a big win with this Invest America program that was just announced as part of the big bill. I know that, Brad, this was your baby and you spent a ton of time on it, but Michael came on board and helped out as well. So why don't you tell everyone the details? We've talked about it before, but tell them the details of what landed.

Yeah. Well, thanks. I remember last Fourth of July we were talking about this bill, and I was sitting right here. Honestly, I thought the chance of getting this passed into legislation was maybe 10% at best, and we had some good fortune.

As you know, the legislation was called the Invest America Act. It was a bipartisan, standalone bill, and it ended up, like a lot of other pieces of legislation, getting subsumed by the reconciliation bill. A lot of these things got packaged together in this one bill, and, of course, it was signed into law on July 4th down at the White House.

I've been at this for 4 years. I tried to get it done under Biden, but the stars just aligned in this moment. Michael was pretty early to get on board and support this, joined the CEO Council for Invest America, and played a critical role with the president to help get it into the reconciliation bill over the course of the last 60 days.

Let's just talk about exactly what it means now that it's become law. I think of this as a pretty significant evolution in the social contract. It creates private investment savings accounts, privately owned for every child at birth, seeded with $1,000 in the S&P 500.

Parents, companies, philanthropists—anybody—can add money to these accounts. You can't take the money out of the accounts; it just compounds in the S&P 500 until you're 18 years old.

We will spend the next year putting the program in place. It has to be launched under the terms of the legislation by July 4th, 2026, the 250th birthday of America.

Basically, we got it expanded so all kids under the age of 18—that's 65 million kids—are eligible. I give a lot of credit to Senator Cruz, who fought to expand the pool of eligibility here. What that means is that they can open up an account, but only children born after January 1, 2025, get the $1,000 from Treasury.

The others will have an account that someone else could put money in on their behalf, and they can add money to it. There are a lot of advantages for their parents adding money and for companies adding money to it.

So it makes a lot of sense, and for Michael and me, I think the key performance indicator here is that if we're having this conversation a year from now, we want to have 50 or 60 million kids signed up. Now, of course, if your child is born after July 4th, then they're going to automatically get an account set up when they get their Social Security number, and they will automatically get the $1,000.

But we have this one-time group—all kids under the age of 18. We're going to have a big campaign to get all those folks signed up over the course of the next year.

And for me, I was reflecting on this over the course of the last few days. At a time when you have an avowed socialist like Zohran Mamdani winning the primary in New York, it seems like the Invest America Act is really just the exact opposite, right? You're both trying to attack the problem of the wealth gap, but this is by getting everybody into the game of capitalism, making everybody actual owners in the upside of America's success, rather than resorting to price controls, attacking businesses and success, and creating more dependency on government.

So I think we're at this critical crossroads in America, and I think the Invest America Act comes at an important point in time. I think a lot of people think of it like a 529 account bill, but I think that dramatically underestimates what this is.

This is a lifetime investment account. They can compound over the course of your life. If you start with $1,000 and you add $750 per year, at 18 that's worth $50,000. At 30, that's worth $170,000, and at 50, it's worth $1 million. So it really is a platform for unlocking dramatic compounding and savings in the upside of capitalism from birth.

And it wouldn't have happened without Michael.

Brad, look, you deserve 99.9% of the credit, so I'm going to give it to you. You really drove this thing passionately for several years, and it's amazing that it got done.

I do think you'll see many companies provide matching contributions. A number of companies have already said they're going to do that, and it'll be like a benefit: come work at our company, have a kid, and your child will get this. It's just going to be super easy for anyone to add to those accounts.

I think it's also a chance to teach every child about financial literacy, capitalism, and free markets. Look up in 15 or 18 years, and you've got 70 million kids with these accounts.

I also think you're going to have philanthropists, and Susan and I will definitely be a part of that, who will say, “Hey, this is a really good way to get money directly to the next generation in a way that is going to compound and make a difference in their life.”

Our foundation has studied this very carefully, and we believe it's worthy of a significant contribution. Brad's been working with the Treasury Department and others to set this up so that any philanthropist would be able to say, “Here's a ZIP code, here's a county, here's a state, here's a group of kids that I'd like to help. I don't know who they are exactly, but I want to help, and I want to help their future.”

I think you'll see a lot of philanthropists get very excited about this. I've had discussions with a number of them, and this could be a major platform for philanthropy in our country.

And just to put a sharper lens on that, Michael, they might back every kid in a state, or everyone in the nation for a year, or just—

Adopt the state. Adopt a series of ZIP codes.

I think, again, it'll be a platform for philanthropy.

Yeah.

Yeah. I think of it, Bill, in some ways like the Giving Pledge 2.0. We've had massive wealth creation in this country—unprecedented wealth creation in this country, right? But one unique feature of America that I don't think there's any other civilization in history that you can point to that has this character is that the super wealthy in America, by and large, want to give away the vast majority of their wealth during their lifetime or shortly after they die. I certainly know that Michael is in that group.

Think about this. In Europe, they invented generation-skipping trusts. It was about coming up with legal mechanisms for creating dynastic wealth so as not to give any of your money away. We have a culture in this country where people want to give away large sums of money.

The challenge is that the charitable infrastructure has not necessarily scaled to meet the needs of people who want to give away billions of dollars at a time. I asked Michael and Susan the question over a year ago. I said, “If you wanted to give away a lot of money in the state of Texas today, how would you do it directly to kids?” There's not a good answer, right? There's not a financial infrastructure in place that has a set of rules associated with it where you could have somebody like the Treasury Department.

We're going to have a pooled Invest America account at Treasury where Michael and Susan, or other philanthropists, could give money to this pooled account, and it would be dispersed to all these kids' accounts subject to all the rules and regulations of use. The kids can't take the money out, but they can see it compound. That simply does not exist today. It's impossible to do that at scale today.

On the long end of the curve, if you think about my family as an example, we do a lot with the East Palo Alto school district and some of these low-income school districts in the state of California, where I can just adopt that school and say, “For every kid in that school, I'm going to give $1,000 a year to.”

So this unlocks, I think, massive creativity around philanthropy. What we know in Silicon Valley is that if you build an open platform, a million applications can bloom. A million ideas can be built on top of this. We've heard from states that want to add $10,000 for every kid born in the state if they graduate from high school in the state.

I think we haven't even scratched the surface of the beautiful competition, the beautiful philanthropy, and the long tail of philanthropy—churches, parents, and friends—that we'll be able to give to these accounts. Our job is to make sure that we make it as frictionless as possible. That's one of the core things that we're doing.

And what you're describing—those accounts don't solely take money at the initiation. They can take money all along the way, which is how you could support a school or something like that.

Correct. So the way it works, Bill, is—and all of this, Michael and I learned a lot about the act of legislation going through this because it's one thing to get it put in the reconciliation bill, and it's one thing to get high-level buy-in—but just in the last 2 weeks, we were negotiating the nitty-gritty. I think this was 23 pages of tax changes in the reconciliation bill associated with the Invest America Act.

Families can give, or recipients can receive, up to $5,000 a year from family, friends, and so on. Companies can give $2,500 a year per recipient tax-free, so pre-tax.

Dell Technologies, for example, has raised its hand and said, “We intend to give to the kids of our employees.” So has Uber. So has NVIDIA. So has Oracle. So has Salesforce. So has T-Mobile. So has iHeartMedia. It's an incredible list that has already come together.

We're going to go to the Business Roundtable. We're going to go to the largest companies in America, and we're going to ask them all to do it. We're not telling them the amount they need to give. All we're saying is, give an amount that's appropriate to your company and to your employee base.

I just heard from Tony Xu yesterday at DoorDash. He retweeted something about this. I heard from Sam Altman over the weekend, once he heard it was passed; he retweeted something about this. I think the business enthusiasm is going to be very big and substantial.

But remember, the most powerful givers are moms and dads, grandparents, friends, birthdays, and bar mitzvahs. All of those dollars generally don't find a home for savings and compounding. We're going to make it as easy as Venmoing in this money, or Apple Paying money in.

One of the studies that we did, which was really profound, in partnership with the Milken Institute found a whole host of things. One was that low-income cohorts tend to save at about the same rate as higher-income cohorts if they have an account. The problem is that nobody in a low-income cohort has a savings account or investment account.

I think you're going to see a lot of giving and a lot of contributions by all sorts of folks once we set this up. We also learned that once we do this, kids are more likely to graduate from high school and college, more likely to start a business, more likely to buy a home, and less likely to be incarcerated. I think the societal ROI of this will be really large over time.

And it sounds like you're going to try and find a way where, if someone wanted to donate—Michael mentioned ZIP codes, but some other way—if people wanted to just target the low-income, most needy, there'll be a way to do that.

Yeah, this was a really important issue to Michael and me. I'll just give you a bit of a window into the weeds. We tried to get household income as one of the targeting mechanisms, and we weren't able to get bipartisan agreement on that. But we were able to get a proxy for that, which is that you can target by ZIP codes.

So, you can target down to groups of 5,000 or more by ZIP code. We think through that geotargeting, so, for example, Michael could target the Rio Grande Valley. I could target East Oakland. There are ZIP codes that you could target that I think certainly include a predominance of lower-income households.

That's fantastic.

Hey, Bill, I know you've been involved in financial literacy and education for a long time. Tell us about the organization you're partnering with and, perhaps, a potential partnership for Invest America.

Yeah, my wife and I have been giving to an organization called Next Gen Personal Finance. There's a gentleman there named Tim Ranzetta who has just been pushing for financial literacy in high schools. We can add a link in here, but from 2021 to 2025, in only a 4-year window, we've gone from 11 states to 29 states. Texas just passed this very recently, a few weeks ago.

The idea—which sounds obvious, but it's actually quite shocking that it's not true—is just to add a semester of financial literacy to the high school curriculum. We send kids out to get jobs, and we haven't taught them how credit cards might take advantage of them, how to build a monthly budget, how to use a checkbook, or how to plan. I think these 2 things complement each other quite a bit, but that's another movement that it's nice to see gaining momentum simultaneously with this one.

Yeah, I talked to Tim. Texas just became, like you said, the 29th state, I think, to require a semester of financial literacy education. Some people said, "The Treasury Department, Invest America—they're not going to own financial literacy." But again, I think when you create a platform of ownership, it makes all of these financial literacy programs and organizations across the country way more effective.

When you're talking to a kid and you say, "Open up your Invest America account on your phone. Let's talk about how you got $12,000–$14,000 into that account. Let's look at how it's compounded. Let's talk about what it means to own the companies that are listed there and what it means to be a shareholder," I think you have a much more engaged student. Today, 95% of those students don't own anything, and they look at their parents and their parents don't really own things. It's a lot harder to get motivated to learn about something when you don't think you're going to have the prospect of ownership.

There are so many great organizations like Tim's out there. I look forward to seeing how they take this platform and run with it to turbocharge their own efforts.

Brad, I know you wanted to mention the budget deficit and the funding for this program and put it in a little bit of perspective, just with all the talk and concern about how big the budget deficit is.

Yeah. Listen, we've had a huge debate among our friend group about this, and some of my friends were even critical that this is part of the problem, if you will. To break this down, the maximum cost of this is $3.7 billion a year. We have 3.7 million kids born every year. If you give them each $1,000, that's $3.7 billion.

Just to put that in context, $3.7 billion is about what we contribute—we give to Afghanistan and Nigeria in terms of foreign aid every year. One of the things as a country we just have to ask is about priorities. Is it more important to give every kid in America a private investment account, a little seed from birth, and get them on the right track, or to give $3.5 billion to Afghanistan and Nigeria? I think those are the types of choices we're going to be forced to make.

I'm not saying that the dollars going to Afghanistan and Nigeria are wasted, but we make these decisions every single day in our budget. For me, that's one angle. The second angle is just as a percentage of our national GDP: this is 1/100th of 1% of our national GDP. It's pretty inconsequential in terms of the overall budget.

The final point on it is, as you've heard me argue, according to the studies that were done on this, this will actually be revenue-contributing 20 to 30 years out, because the taxes you have to pay when you exit the accounts on the capital gains will be more than what the government is contributing on an annual basis into the accounts. Among the things we should be worried about when it comes to the budget, I don't think this is one of them.

3. Government Spending and Budget Deficit

However, I would say unquestionably that I remain as concerned about the budget deficit as ever. I've been a supporter of a balanced-budget amendment for a long time. I happen to think that this is something that is aligned with that, not at odds with that. Making every kid a capitalist from birth is going to better align us with the policies that allow the country to continue to grow, and I think growth is a critical element to making sure that we get our deficit-to-GDP back in a manageable place. Michael, I know you care a lot about that issue. Any other thoughts on that particular point?

Yeah, government's obviously been spending too much, and there's been some renewed attention and focus on that. That's a good thing. It gets priced into the currency, right? We see it in all the effects, whether it's inflation or the value of the currency, and you can't really escape that.

I think the spending has to come under control. Now maybe we get this incredible productivity lift. I'm sure we're going to talk about that as we get to the AI fund portion here. But we shouldn't be spending so much more than we're taking in as a government. I've sort of stepped back from the hysterics, and you say we don't have a loan-to-value problem as a country. We have a spending problem.

I want to dig into that because it's a really important point. You talked about loan-to-value. When you say loan-to-value, what do you mean by that?

Loan-to-value is a common term and phraseology used in banking and credit markets, and essentially refers to the amount of a loan relative to the value that it's being borrowed upon. If you think about the deficit as against the value of all the assets in the United States, we don't have a loan-to-value problem.

And, of course, the total value of all the assets in the United States is a couple hundred trillion. Our annual deficit is $2 trillion. So you would look at that and say, as a loan-to-value, that's not an issue at all.

Well, I would look at the total deficit as against the total value of the assets.

So, that's $36 trillion of debt against $200-plus trillion of assets.

Right. Now you have to take into account private assets and private debt also, so it would be a different equation there. But the government also has taxing authority, and so it could increase the taxes. Net it all out: the government shouldn't be spending what it's spending relative to what it takes in. There are many ways to address that, but we should be worried about where the deficit is and the rate of increase.

Let me ask a question about that. First, maybe to level-set: the argument out of the White House is that the reconciliation bill cuts the deficit. The deficit was about $1.9 trillion. Their argument is that it cuts the deficit by about $150 billion a year, so $1.5 trillion over 10 years. They also argue you get another $250 billion in tariff revenue, incremental from the start of the year. We saw that in the run-rate revenue in the month of May.

You add those 2 things together, and now you're at $400 billion. So, if you were at a $1.9 trillion deficit, now you're down to about a $1.5 trillion deficit. By my math, that drops it to about 5% deficit-to-GDP. Bessent has said he will get it to 3% deficit-to-GDP, which is what most people say is reasonably healthy. I think people would like to not have any at all, but I think most people view 2% to 3% as reasonable.

He thinks he can get there by 2027 or 2028 through the 2 things I just mentioned: tariff revenue and the deficit reduction in the reconciliation bill, plus an incremental 100 to 200 basis points of growth in the country caused by lower taxes, less regulation, AI productivity, and so forth. Is your view that we just have to wait and see? Does that show up, or does it not show up?

Well, obviously, we have to wait and see. On the trade and tariffs front, I think this is very tricky, right? We have products flowing back and forth, and we have services flowing back and forth. If you think about the market cap of U.S. companies versus the rest of the world, the U.S. is doing really well relative to the rest of the world in market cap. The reason is that we have a substantial lead in the most valuable industries in the world.

Correct.

The issue there is that if you think about the trade in products, you also have to think about the trade in services and how that's going to be dealt with in a negotiation. I don't know how that'll all get sorted out, but I don't think it's a simple one-line-item fix.

Right. Right. No, I think it's all relevant right now.

Elon’s talking about forming a third party, the America Party, really in response to what appears to be frustration over DOGE and the budget deficit, and concerns by folks like Ray Dalio about a debt spiral in the United States. You’ve got guys like Scott Bessent saying, “Elon, you catch rockets; leave the finances to me.” Bessent seems very confident that he’s going to get this back down to a 2% to 3% deficit to GDP.

I actually like the suggestion, Bill, by the way. Rather than forming a third party, which seems to me just chaotic and a lot of overhead and has not historically been that successful, I would love to see Elon—if this is his main issue, if it’s the budget deficit and debt, which I would love to see him take on—do a series of things.

Number 1, he could really sponsor a balanced-budget amendment to the Constitution of the United States under Article V. If he put $10 billion against that effort, it would be the single largest constitutional effort in the history of the country. I think there’s broad bipartisan support for a balanced-budget amendment. We have 32 states that have supported this in the past. I think you only need 34 to get a constitutional convention called and 38 states to get it ratified.

It hasn’t happened. The founders made it hard to amend the Constitution for a reason, but I actually think if he put those types of dollars and that type of focus behind it, we could get it done. And then on top of that, he could target both Democrats and Republicans in primaries around this issue.

To me, it just seems like that targeted approach, that very focused approach to balancing the budget, would have all sorts of positive impacts. Number 1, it keeps the country focused on this issue. It keeps this administration focused on this issue. I think you have an outside chance at getting a constitutional amendment, and you certainly are going to have a lot of Republicans and Democrats who will run on that issue if they think they’ll get Elon’s support.

I’m not sure how this will all evolve, whether there’s going to be a third political party or not, but I would love to see this issue get dealt with. I remember Ross Perot tackling it in 1992 and 1996. Michael, I know you knew Ross, and to me, that type of attention is the type of attention that we’re going to need.

4. The AI Talent War

Why don’t we shift gears here for a second? This is one I’ve been dying to ask you both about. There’s this really unprecedented war for AI talent going on. It was kicked off by Zuckerberg and Meta. They made the acqui-hire of Scale AI for $15 billion. They brought on board Alexandr Wang to help lead that effort. Then they brought on board Nat Friedman and Daniel Gross. They’ve poached a bunch of people from OpenAI, a bunch of people from Google, and now, today, another announcement of somebody from Apple.

The talk is $75 million to $100 million annual pay packages, massive signing bonuses—really dollar amounts, Michael and Bill, that I don’t think I’ve ever heard of in the tech industry. So, Bill, given that recent set of facts, what is this? Is this a good thing? Is this a bad thing? What do you think the downstream implications of this are?

Well, I would back up a little. I don’t think it started with Meta. I think it started with the cycle that we’ve been under in the private funding market. We saw some of this stuff during what was likely ZIRP, but we’ve moved to a world—and I talk about this in detail on O’Shaughnessy’s podcast from a few weeks back—where, when there’s a successful company, the late-stage private market at large tries to shovel-feed cash into it.

So we have private companies that have raised not just $100 million but $1 billion or more. And we have a handful of private companies, including OpenAI, that are voracious and audacious enough to burn $2 billion, $3 billion, $4 billion, $5 billion a year.

You start doing that, and you create a situation where private companies have an odd advantage over public companies in that their investors are more willing to let them lose a lot of money than public investors may be willing to. And so they get bold and audacious.

OpenAI and Anthropic were paying people tons of money before Meta did this. They were paying them $10 million a year—maybe smaller than what you were talking about, but they were doing it. They were providing liquidity earlier, like 2 years in instead of waiting for 4, and providing liquidity as a private company. All of these things, in some cases, may have let these people leave because they didn’t have any lock-in.

That part may have backfired. But in Zuck, you have someone who’s had his back against the wall a couple of times, gotten bold, changed what he was doing, and succeeded again. So he has the conviction that he’s willing to take a big bet.

I think he’s very willing to look at cost as a percentage of his market cap and to view risk as spending against a percentage of his market cap. Not everyone’s capable of doing that. I think it may be the right math, actually, in terms of how big a bet he wants to make.

What he has done here in the past 3 weeks is an experiment that’s never been tried before. But there’s unlimited free agency in business, unlike sports. And he just went and bought the 27 Yankees of AI.

Yeah, and I think your point is a great one. And listen, we’re shareholders in Meta. We’re shareholders in OpenAI. I wouldn’t be a shareholder in Meta if I didn’t think—in fact, I remember back in 2022, when we took our big position there and people said to me, “Oh, what are you doing? This is a founder-controlled company. He’s never going to become more efficient. He’s never going to do these things.”

I said, “The whole reason I want to be all-in on this company is that it’s founder-controlled. I think it is a massive advantage that he has today.” And he’s talking about risking 1% of his company in order to reboot around AI. That seems to me to be a very, very rational economic decision.

And this is just a talent war. He’s got to—Llama 4 was not where it needed to be to compete head-to-head, but he has one advantage none of those other companies have. He has the world’s biggest printing press shooting out billion-dollar bills. He’s not relying on the beneficence of venture capitalists. The guy has a business model that is generating the cash to fund all this.

So he’s leveraging that cash as a source of competitive advantage, which seems to me to make a lot of sense. I think it’s going to make it very difficult, and that’s why I was asking about the downstream implications, Bill. If you’re a company that’s trying to compete against that, I don’t think many venture companies can compete against that on a durable, long-term basis.

Certainly not the real startups.

Yeah. I was having a discussion with a real AI startup founder this weekend, and he was asking about talent. I don’t know what you do. I don’t think you hire anyone who’s in the top 1,000 in the Bay Area. You won’t be able to afford them.

But I do think there’s a fundamental question because it’s easy to say “the percentage of market cap” and make that bold decision. There are cultural implications, right, of bringing in employees who make radically different amounts of money from the rest of the employee base. How do you think that will be? How difficult will that be to manage?

I think it’ll be a challenge culturally, for sure. He could have a long line outside of his door, with people wanting this or complaining about that, and that could be a distraction.

I think people generally have a sense of fairness, right? They want to be treated fairly relative to others and relative to the opportunities that they have out there in the overall market. And if they feel that they’re not being treated fairly, that’s going to be a problem. So I don’t know how that gets sorted out.

I do think the math could work for them, given everything you guys just talked about. And obviously, if you reduce this down to a race to superintelligence or something along those lines, the size of the prize is tremendous. They do have an incredible business that is aided by these advancements in a big way. And there aren’t a whole ton of companies that can go do this.

Yeah. And by the way, Brad, you mentioned that they have this unfair advantage of this huge printing press, but Apple and Google have the same exact printing press and chose not to do this.

Yeah. But neither of them are controlled by founders, and that’s the point I was trying to make. These are the types of bets that I think are very, very difficult for a Google or an Apple to make, for the reasons you mentioned, Bill.

Can they sell it to the public markets? Do they have the type of decision-making in the boardroom that allows this to occur? At the end of the day, I think at Meta, if Zuckerberg wants to do it, that’s what’s happening. And that board gets on board.

In fact, he’s reshaped the board over the course of the last couple of years with folks who are, I think, signed up for this mission with him.

Michael, to your point, that's why I think he reorganized this into the kind of superintelligence division. I think the way they'll try to manage this culturally is to say, "Listen, there's going to be an elite SEAL Team 6 group, which is called superintelligence, and we're going to pay them elite pay because it's good for the entire business. That doesn't mean we're going to inflate everybody else."

In fact, what I think Meta will do is you'll probably see them rolling back, like you see with Microsoft and Amazon. My sense is that companies are generally going to get smaller on the backs of the productivity gains from AI, but they'll redeploy some of those profits into these areas. If you're in the model business and you want to be on the frontier, competing for superintelligence—and there are only 5 to 7 companies that really are in that game—then I think you're going to have to have something similar.

In the case of OpenAI, it's only 2,800 employees, and they're all part of that division, effectively. But you have to really get scale quickly, because if you're not bringing in $10, $20, $30, or $40 billion of annual revenue, I don't think you can stay in this game. The question is whether Anthropic, xAI, and OpenAI have sufficient escape velocity that they can take on this frontal assault by Meta and still compete. My sense is OpenAI does. My sense is both of those companies do, but it's not a long list that can compete with that.

And by the way, the Nat Friedman addition was particularly interesting, just with his GitHub background. We had talked in the past about Meta making a couple of hires on the enterprise side, and we'd heard rumors of certain payments when they passed through the cap on the open models. You have to wonder, with Nat coming on board, if there are more aspirations on the enterprise side.

5. AI's Role in Economic Growth and Productivity

Yeah, it's a great point, and certainly it creates some optionality there. Hey, Michael, a question for you. On this related topic of productivity gains from leveraging AI and what you're seeing at Dell, we've talked on this podcast about what we call the golden age of margin expansion—the idea that you're seeing AI reaccelerate your top line in a pretty dramatic way, but at a lot of companies at the same time, you're able to do more with less.

Is that overstated, or do you think that we're in this phase over the next 3, 4, or 5 years where, generally, as an economy and certainly within a lot of companies, they're going to be able to have their top lines grow faster than their operating costs because of AI?

It's absolutely real, Brad, and we are doing it. We know of other companies that are doing it. Maybe only 10% of large companies have figured this out, and the other 90% are a bit confused at this point.

If I step back and look at this, 10% productivity improvements are pretty easy. Twenty percent is reasonably common; sightings of 30% or 40% are massive numbers. You have a $114 trillion global economy in 2025, and the services economy is two-thirds of that. If we believe that a 10% improvement is possible, and you keep it simple and say a 10% improvement, that's worth $10 trillion.

Yeah, and that's not where we are.

It's a lot less than that. I don't want to get too ahead of myself here, but I do think there is a big change occurring. We're just at the beginning of it, and it's going to affect every part of our world.

Well, you have particular standing here, Michael. You saw the productivity gain that came from a computer on every desktop.

That was the '90s. We were talking about the '90s. That was fun.

You saw the productivity gain from the internet. Now you're 2 years into observing this. Are the productivity gains from this going to be as big or bigger than what we saw from personal computers and the internet?

Oh, it's far bigger. It's far bigger. I feel 98% confident that it's far bigger than the PC.

The internet. What about the internet?

Yeah. Of course, all these build on each other, right? It's compounding, but this is bigger because it is essentially all knowledge work. I think it's an expansion of the pie. It's really easy to figure out what will be more efficient and how you can reduce costs, but if you go back 20 years ago, it was very hard to see where the new jobs would be.

I think we're in a similar situation here as well. I do think it will be expansionary for the overall economy and for prosperity, well-being, and human potential broadly across all domains, whether it's education, health, societal outcomes, et cetera. But this feels bigger.

I remember in 2001 and 2002, some companies that were early to Google figured out how to game the internet and Google to build a giant business. Booking.com, for example, figured out how to arbitrage the internet and Google to build this giant business. They figured out productivity gains before the next person, and that was hugely advantageous.

When you say that only 10% of companies are leveraging this today, it sounds like the same thing: The early companies are really there, but there's a huge amount yet to come.

When I think about the big companies in the world, I'm talking about $10 billion-plus-revenue companies, $50 billion- or $100 billion-plus-revenue companies. These companies have an incumbency of sorts. They have data, customers, brands, intellectual property, and so on.

But if they don't move quickly to reimagine their businesses given all this technology, they will be destroyed by new companies that come in with a totally clean slate. You can already see signs of that happening. I think this is all going to play out in the next 3 to 5 years, and it will become an urgent priority for companies to reimagine themselves.

What we've done at Dell is that our team knows this because we talk about it all the time internally. Almost 2 years ago, I stood up in front of a group of our leaders and said that 5 years from then—which would be 3 years from now—we were going to have a new competitor. That new competitor was going to be in every business that we're in, except they were going to be faster, more efficient, and more capable, and they were going to put us out of business.

The only way we're going to prevent that is to become that company. This is how we're going to do it. I laid out our best guess as to how to do that 2 years ago. We're pretty far into that path and well on our way, and it's working, but it's not an easy thing to do. This is gut-wrenching stuff—to reinvent and reimagine. We've had to do it many times. If you don't do it, you just go out of business.

And that's no fun. So we're not doing that. Not everybody wants to do it. It's true. It's hard.

6. AI Compute–Explosion in Demand

Yeah. I was wondering if you could expand on that a little bit. Your server division is your fastest-growing division. We've talked on this podcast about some of the big wins you've had as part of large AI clusters. How did you get Dell in a position to be part of that next wave, and what's the key value add from your products in those large deployments?

Last year, our server and networking business grew 58% year over year. In the first quarter, we received $12.1 billion in AI orders. By the way, our shipments for all of last year in AI servers were about $10 billion. Two years ago, we were at $2 billion.

Yeah, it was not very much 2 years ago.

We took orders in the first quarter for over $12 billion, and last year we shipped about $10 billion. This is growing super fast, and now we have a backlog of a little over $14 billion.

What happened? We're already the leader in servers. We saw the GPU opportunity, and it's a combination of things. When NVIDIA releases a reference design, it's kind of a reference design. It doesn't really work. We love NVIDIA, but somebody's got to make all this stuff.

There's a ton of engineering involved, and obviously there's the logistics and the supply chain. Building these 100,000-plus-GPU clusters and making them work reliably is super complex. It's a combination of engineering and operations. We often will help with the financing of these through Dell Financial Services.

The scale of these things is enormous. We talked about this at Dell Technologies World. Right now, we're deploying systems that will deliver more than 50 trillion tokens per month. If you put that in the context of Google's or Microsoft's statements, this is massive scale—massive-scale systems.

I don't think there are a ton of companies that are able to do this and have them work reliably.

And Jensen has said that you guys have distinguished yourselves against your competitors, other ODMs like Foxconn or Quanta. You've been first to market. You're launching the GB300s right now. You're a partner.

Yeah, we delivered the first GB300s a couple of days ago to CoreWeave. We announced that we actually delivered another GB300 system to another customer. I don't think we've disclosed who that is yet, but informed listeners of this podcast will probably guess.

A year ago, we were all sitting around and talking about the ups and downs of the overbuild in 2000 around the internet. Yet when I look at the trajectory we're on, I saw Mike Intrator on CNBC today, and he said, “Listen, we're still underestimating the amount of demand that's out there in the world.”

When he says it, or when you say it, or when Jensen says it, in some ways people would argue it's self-serving. Of course, you guys are going to say that. That's your business. You're going to tell everybody your business is great. But you're known as a very sober guy who tells it like it is.

What I want you to do is reflect a little bit on the comparison between this and the period in early 2000 when we did get overbuilt. As the saying goes, every shortage ends up in a glut. Why are we not near that point yet today in this market?

Well, you guys, as students of the market, can go back and review what the multiples were on earnings and cash flow around that time. We're nowhere near that, for the most part.

But if we go back to the underlying activity here, it's all about the tokens, right? And as we go from basic queries to test-time compute to deep reasoning to agents and multi-agent systems, the number of tokens just explodes. What are we talking about in tokens? We're talking about knowledge, right?

Exactly. I don't know about you, but I'm using these tools like 50 times a day as my thought partner to solve problems and quell my curiosity. My usage is skyrocketing, and I'm often using multiple models. It's going out there and querying all these websites, doing calculations for me, and helping me solve problems faster than I ever could in the past.

This, I think, is just at the beginning, right? The substrate for all of this, of course, is compute and data, right? So we love that at Dell Technologies because that's what we do. There's just a ton of growth here. I think it will also be highly distributed. It'll occur in devices. It'll occur at the edge. It'll occur in all sorts of places, and it does feel like we're still a lot closer to the beginning here.

What can you share about on-prem AI deployments? Michael, are you seeing anything interesting there?

Yeah. In the last year, we delivered a little over 3,000 of these Dell AI Factories. Those are increasingly going to enterprise and commercial customers that want to bring the AI to their data, not the data to the AI.

There's just a ton of data that's still on-prem and being generated on-prem. It turns out these large models are fantastic, but you don't always need the largest model to solve every problem. A lot of the corporate use cases are perfectly done with smaller models and open-source models.

So you see this enormous proliferation on Hugging Face of models of all shapes and sizes, with tons of cascading innovations. I think this is going to be highly distributed. We're definitely seeing growth in on-prem, and colos are also a big deal because many customers don't want to have the data center themselves. They'll put it in a colocation facility, and we've also adopted the consumption-type model, so you can pay on a usage-type basis.

What is your view when you look at just the relative distribution between the custom ASIC world and what you see happening across folks like AMD and Nvidia? You have an interesting perspective both as a consumer of these products and as a builder and distributor of the products. Are there any pending big changes? As you look ahead over the course of the next year or 2—that's probably as far ahead as you can see—does it look like the relative landscape is stable, or are there big breakthroughs coming that may unseat somebody like Nvidia?

Nvidia is in a great spot. To your question, I think for the larger model companies and hyperscalers, certainly custom ASICs are gaining a lot of share. When you have control over the workload and you can take the time to optimize your workload, that's certainly going to be a part of what occurs in the infrastructure.

But it's not a large number of customers. You were talking about the number of companies developing models. It's sort of that number of customers, but they're large, as you've seen with Google and Meta and others who are deploying the ASICs.

7. Market Check

Maybe, just to be respectful of time, Bill, I could talk to Michael for hours about this particular subject. But one of the people I talk to when the market's going wild is Michael. We certainly saw that earlier this year. It's pretty incredible to see the snapback that we've seen out of the Nasdaq and the S&P. I think the Nasdaq is now up 32% off of its bottom 2 months ago.

Just as a data point, I think Dell got as low as—I'm not sure—$72. It's back at $120 now. That is an incredible bounce off the bottom, but it's still basically up, I don't know, 5 or 10% on the year. It's not like it's in this astronomical range when you look at year to date or over the course of the last 12 months.

Here we are. We have the Nasdaq, the Qs, and the S&P at all-time highs. Bitcoin's near an all-time high. The VIX is back to 15 or 16, basically where it was in February, despite all of the things around tariffs. The 10-year—everybody talks about the great debt spiral that we have in the country—but the 10-year has been between 3.7 and 4.7% for the last 2 years. It's at 4.2%, smack dab in the middle, if not at the lower end of that range.

You see companies like TSM, Nvidia, Microsoft, Oracle, Booking.com, Uber, and DoorDash at all-time highs. Notwithstanding the fact that they're at all-time highs, you have Tesla down over 20% on the year, Apple down 15% on the year, Google's down on the year, and Amazon basically flat on the year. So you have a lot of dispersion in the market.

Does it feel to you, again, like we're in this bubble territory? Set aside your stock. I don't want you to comment on your stock. I'm just talking about the market at large. Are the U.S. markets higher in 3 to 5 years, or are they not, given where we sit today?

I would bet they're higher. I would bet that more and more companies figure out how to grow their businesses. I talked earlier about productivity and efficiency. I think the ultimate benefit is going to come from the speed at which companies transform and the growth that they're able to create.

That's certainly how we see it in our business. I think a lot of these companies will be able to compound their earnings on a double-digit basis, and the market, largely—the overall indices—will become more valuable.

Yeah, I do think that at this moment in time we're seeing a lot of dispersion. I mentioned it, right? Some companies are down this year, some companies are up a lot. I really think the companies that are leveraging AI, that are in a position to leverage it and to capture that margin expansion, are going to see a reacceleration.

We've heard this out of folks like McDermott, Sridhar, and Jensen at all these companies—how they're reaccelerating top-line growth, but they're not adding people, right? It's kind of net flat. We see this out of Uber. We see it out of Dell.

When the market dipped down, our share buyback program went into high gear. We bought back 22 million shares.

Right. The stock worked out well for you.

As I look at this flight path, we just landed the reconciliation bill. There was a lot of uncertainty in the world to start the year. One question was what was going to happen? Was this reconciliation bill going to pass? Now it's passed, so we have tax predictability, right?

You have an extension of the existing tax regime, and then you have no tax on tips and no tax on overtime. So you have this incremental stimulus now coming from the reconciliation bill. On top of that, tariffs, while still up in the air, the market has kind of digested the tariff stuff, right?

Absent some big blowup between us and China, if we follow the Bessent accords that they reached in Switzerland and then reiterated in London, it seems like the big pieces of the tariff puzzle are falling in place. On the rate front, the market is estimating that the next move is down. Whether we're going to have 1 or 2 rate cuts at the end of the year is the question. The Fed is saying we're on hold; we're going to wait and see whether or not inflation reaccelerates this summer due to tariffs.

So, that’s what everybody’s eyes are on over the course of the next 6 to 8 weeks. Does core PCE tick up due to those tariffs? I’m taking the under on that, but we’re going to have to wait and see.

And then on fundamentals, I think what we’re hearing from companies—and this is where the rubber meets the road—is earnings. I think we had 85% of companies beat in the S&P 500 in the quarter. And if you just go through and look at keywords, it was “accelerating,” it was “AI,” it was “reinventing our business.” There is a real growth feeling in the market and among these companies.

And so, from our perspective—and we try to give people an indication of where we are—I was as negative, as you well know, Michael, early in the year as I’ve been in 10 years, because I thought if we were going down the path of Navarro and $2 trillion of tariffs, all bets were off. That was a scary path.

We talked about how, if they went down that path, I thought they would reverse course because it wouldn’t work. This, I think, is a very iterative team that will experiment and lay some stuff out there. Not all of it’s going to work. There’ll be some bad ideas, and then they’ll reverse course.

Yeah. I hope we don’t. We did do that. I hope we don’t snatch defeat from the jaws of victory here with policy, though. Going back into the tariff game, some type of bold confrontation with China, and our AI policy—I mean, one thing we didn’t talk about this past week is that the AI moratorium got removed from the bill, and we’re going to have 70 state laws in the United States, which is not great for AI startups.

So, anyway, I hope bad policy doesn’t upset what would otherwise be a very potent landscape based on AI.

Yeah, fully agree. I think it’s one thing that the 3 of us are in violent agreement on. One of the things that’s moved this country forward for the last 3 decades is that we’ve led globally in technology. And we’ve led globally in technology because we’ve allowed our best technologies to move freely around the world, compete, and win.

This is the first time since I’ve been in this business that we’re talking about export controls and AI diffusion laws that are restricting the ability of our technology to go compete and win. There’s both the question mark as it relates to inside of China, but also the question mark outside of China.

And while we’ve seen the repeal of the Biden AI Diffusion Rule, what I’m told is that no new licenses have been granted for distribution of AI technologies around the world, despite all of the discussion around this. So, it’s critical that Washington follows through and that we accelerate diffusion around the world of the entire American AI stack, and that we don’t regulate that out of Washington.

I think there are some legitimate regulations that you can have as it pertains to China. But even there, I would much rather let our deprecated chips out of NVIDIA go compete against Huawei in China. Keep the developer mindshare in China, because it’s going to make it easier for us to win globally and elsewhere around the world.

And I think it’s important that Michael, myself, everybody else—Bill, you—all those voices are being heard. We’re not out of the woods on this by a long shot.

By the way, you reminded me of one other thing I’d just like to harp on, which is the skilled immigration piece. Someone highlighted to me that they made this huge wanted poster of all the people that Meta has borrowed from other companies, and 60% or 70% of them were of Chinese origin.

As I understand it right now, there are Ph.D. students or candidates in China who can’t get visas to enter the United States. We go back to what Trump said on All-In, that he wanted to staple a visa to every diploma. I’d really like to get that conversation going again. It would be very powerful for the country to increase skilled immigration, and it feels like we might be decreasing it.

Yeah, absolutely agree. And to your point, Brad, if we don’t aggressively work to sell our technologies around the world, other countries are going to do that.

I’m reminded of a story from a long time ago. The Defense Department had this thing called MTOPS. It was a restriction on how fast the computer was, and you had to get approval from the government to sell it.

I was in this group of technology CEOs, and we went to the Pentagon to talk to the generals. Before we went, we went to a Toys “R” Us store and bought a PlayStation. We took it out of the box and brought it to the Pentagon, to this big room. We set the PlayStation down and said, “This exceeds the MTOPS restriction, right?”

“But unfortunately, it’s made by a Japanese company, so it doesn’t fall under the rules. Anybody can buy it. It’s also $399, right? So, if you think you’re going to control access to little things that move easily, you’re kind of fooling yourselves.”

Yeah.

And so, we have to come up with more intelligent ways to restrict access to the most advanced technologies. Oftentimes, you just get all kinds of unintended consequences with these rules that are created, and they don’t create the outcome that the government was originally looking for.

Well, I think that’s a good way to wrap. Michael, it’s awesome having you here. Michael and myself, Dara Khosrowshahi, David Solomon from Goldman Sachs, René Haas from Arm, Bill McDermott from ServiceNow, and a group of us were at the White House a few weeks ago to testify on the Invest America Act, and Michael kicked it off and captivated the entire room. He reminded everybody:

We view this initiative as a powerful platform for philanthropic innovation aimed at helping children thrive wherever they come from, particularly those families who have been historically left behind.

Mr. President, you articulated it perfectly. These Invest America accounts will give every new American child a genuine opportunity to participate in history’s greatest engine of economic growth: the American economy.

The funds in these accounts, invested in American enterprise and innovation, will grow over time into substantial nest eggs, providing support for education, homeownership, and starting families. The ability of families, friends, benefactors, and employers to match the government’s generosity amplifies the life-changing potential of this initiative.

Thank you, Mr. President, for your visionary leadership on this critical issue. These Invest America accounts will profoundly impact countless young Americans, ensuring they truly benefit from what Abraham Lincoln described as the right of every American—the right to rise.

As I sit here on the Fourth of July weekend, I’m super grateful to you, Michael. You did a huge service to the country by helping us get the Invest America Act passed.

And I think everything that we just talked about here, including allowing American technologies to go compete—remember, these Invest America accounts are only worth something if America does great, right? The fact of the matter is Warren Buffett has said the smartest thing he did was just bet on America. He bet on America.

I’m betting that the next 50 years, the next 100 years, are going to be an American century again. But we can’t get in the way of the innovation and the entrepreneurship and the creative destruction, frankly, that has allowed America to be so great.

Finding that balance between government and Silicon Valley has always been challenging, as you just related, Michael, with MTOPS. But we have to show up. We have to have a voice. We have to continue to push in that direction.

I think if we’re allowed to compete, our best days lie ahead. If we get in the way, Bill, like you talk about, then I think we can upend our advantage. Thanks for joining us, Michael. Great to see you.

Appreciate it.

Great to see you. Bye-bye.

We’ll talk soon. Take care.